1/* $NetBSD: n_cabs.S,v 1.7 2014/10/10 20:58:09 martin Exp $ */ 2/* 3 * Copyright (c) 1985, 1993 4 * The Regents of the University of California. All rights reserved. 5 * 6 * Redistribution and use in source and binary forms, with or without 7 * modification, are permitted provided that the following conditions 8 * are met: 9 * 1. Redistributions of source code must retain the above copyright 10 * notice, this list of conditions and the following disclaimer. 11 * 2. Redistributions in binary form must reproduce the above copyright 12 * notice, this list of conditions and the following disclaimer in the 13 * documentation and/or other materials provided with the distribution. 14 * 3. Neither the name of the University nor the names of its contributors 15 * may be used to endorse or promote products derived from this software 16 * without specific prior written permission. 17 * 18 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 19 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 20 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 21 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 22 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 23 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 24 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 25 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 26 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 27 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 28 * SUCH DAMAGE. 29 * 30 * @(#)cabs.s 8.1 (Berkeley) 6/4/93 31 */ 32 33#include <machine/asm.h> 34 35 .globl _C_LABEL(__libm_dsqrt_r5) 36/* 37 * double precision complex absolute value 38 * CABS by W. Kahan, 9/7/80. 39 * Revised for reserved operands by E. LeBlanc, 8/18/82 40 * argument for complex absolute value by reference, *4(%ap) 41 * argument for cabs and hypot (C fcns) by value, 4(%ap) 42 * output is in %r0:%r1 (error less than 0.86 ulps) 43 */ 44 45/* entry for c functions cabs and hypot */ 46#ifdef WEAK_ALIAS 47WEAK_ALIAS(hypotf, _hypotf) 48#endif 49 50ENTRY(_hypotf, 0) 51 cvtfd 4(%ap),-(%sp) 52 calls $2,_C_LABEL(_hypot) 53 cvtdf %r0,%r0 54 ret 55 56#ifdef WEAK_ALIAS 57WEAK_ALIAS(hypot, _hypot) 58WEAK_ALIAS(hypotl, _hypot) 59WEAK_ALIAS(_hypotl, _hypot) 60#endif 61 62ALTENTRY(cabs) 63ENTRY(_hypot, 0x8040) # save %r6, enable floating overflow 64 movq 4(%ap),%r0 # %r0:1 = x 65 movq 12(%ap),%r2 # %r2:3 = y 66 jbr cabs2 67 68/* entry for Fortran use, call by: d = abs(z) */ 69ENTRY(z_abs, 0x8040) # save %r6, enable floating overflow 70 movl 4(%ap),%r2 # indirect addressing is necessary here 71 movq (%r2)+,%r0 # %r0:1 = x 72 movq (%r2),%r2 # %r2:3 = y 73 74cabs2: 75 bicw3 $0x7f,%r0,%r4 # %r4 has signed biased exp of x 76 cmpw $0x8000,%r4 77 jeql return # x is a reserved operand, so return it 78 bicw3 $0x7f,%r2,%r5 # %r5 has signed biased exp of y 79 cmpw $0x8000,%r5 80 jneq cont /* y isn't a reserved operand */ 81 movq %r2,%r0 /* return y if it's reserved */ 82 ret 83 84cont: 85 bsbb regs_set # %r0:1 = dsqrt(x^2+y^2)/2^%r6 86 addw2 %r6,%r0 # unscaled cdabs in %r0:1 87 jvc return # unless it overflows 88 subw2 $0x80,%r0 # halve %r0 to get meaningful overflow 89 addd2 %r0,%r0 # overflow; %r0 is half of true abs value 90return: 91 ret 92 93ENTRY(__libm_cdabs_r6,0) # ENTRY POINT for cdsqrt 94 # calculates a scaled (factor in %r6) 95 # complex absolute value 96 97 movq (%r4)+,%r0 # %r0:%r1 = x via indirect addressing 98 movq (%r4),%r2 # %r2:%r3 = y via indirect addressing 99 100 bicw3 $0x7f,%r0,%r5 # %r5 has signed biased exp of x 101 cmpw $0x8000,%r5 102 jeql cdreserved # x is a reserved operand 103 bicw3 $0x7f,%r2,%r5 # %r5 has signed biased exp of y 104 cmpw $0x8000,%r5 105 jneq regs_set /* y isn't a reserved operand either? */ 106 107cdreserved: 108 movl *4(%ap),%r4 # %r4 -> (u,v), if x or y is reserved 109 movq %r0,(%r4)+ # copy u and v as is and return 110 movq %r2,(%r4) # (again addressing is indirect) 111 ret 112 113regs_set: 114 bicw2 $0x8000,%r0 # %r0:%r1 = dabs(x) 115 bicw2 $0x8000,%r2 # %r2:%r3 = dabs(y) 116 cmpw %r0,%r2 117 jgeq ordered 118 movq %r0,%r4 119 movq %r2,%r0 120 movq %r4,%r2 # force y's exp <= x's exp 121ordered: 122 bicw3 $0x7f,%r0,%r6 # %r6 = exponent(x) + bias(129) 123 jeql retsb # if x = y = 0 then cdabs(x,y) = 0 124 subw2 $0x4780,%r6 # %r6 = exponent(x) - 14 125 subw2 %r6,%r0 # 2^14 <= scaled x < 2^15 126 bitw $0xff80,%r2 127 jeql retsb # if y = 0 return dabs(x) 128 subw2 %r6,%r2 129 cmpw $0x3780,%r2 # if scaled y < 2^-18 130 jgtr retsb # return dabs(x) 131 emodd %r0,$0,%r0,%r4,%r0 # %r4 + %r0:1 = scaled x^2 132 emodd %r2,$0,%r2,%r5,%r2 # %r5 + %r2:3 = scaled y^2 133 addd2 %r2,%r0 134 addl2 %r5,%r4 135 cvtld %r4,%r2 136 addd2 %r2,%r0 # %r0:1 = scaled x^2 + y^2 137 jmp _C_LABEL(__libm_dsqrt_r5)+2 138 # %r0:1 = dsqrt(x^2+y^2)/2^%r6 139retsb: 140 rsb # error < 0.86 ulp 141